Automated extraction of chest CT derived body composition biomarkers of tissue morphology and texture.
Principal Investigator
Name
Miranda Kirby
Degrees
PhD
Institution
Toronto Metropolitan University
Position Title
Associate Professor
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-1095
Initial CDAS Request Approval
Jul 18, 2023
Title
Automated extraction of chest CT derived body composition biomarkers of tissue morphology and texture.
Summary
Chest-CT derived measures of body composition, such as skeletal muscle and adipose tissue areas, have been used to predict patient outcomes in debilitating lung diseases such as COVID-19, interstitial lung disease, chronic obstructive pulmonary disease, and lung cancer. However, most studies use a manual pipeline to perform the tissue segmentations, which is both time consuming and introduces observer variability. This project will propose an automated analysis to extract the skeletal muscle and adipose tissue from chest CTs. In addition, biomarkers of tissue morphology and texture will be derived using state-of-the-art radiomics analysis and shape descriptors. Finally, the ability for the biomarkers to serve as predicters for patient outcomes in lung cancer will be evaluated.
Aims
1. To segment the cross-sectional skeletal muscle and adipose tissue from chest CT scans.
2. To extract descriptive biomarkers of tissue morphology and texture from the segmented structures.
3. To evaluate the predictive capabilities of the biomarkers with lung cancer outcomes.
Collaborators
None.